Robotic Process Automation (RPA) is extremely popular in banks and insurance companies, where software robots take over repetitive steps of business processes such as account opening, data validation, or KYC/AML checks. As these processes grow more complex and increasingly rely on machine learning, controlling the RPA computations becomes difficult, and operating the RPA infrastructure (servers, licenses, …) becomes costly.
Together with Pointee Interactive s.r.o., we are developing the Scheduler, a software module that plans the runs of RPA robots and other automations (currently at TRL 4). The main goals of the project are:
Automatic generation of the Scheduler inputs from historical data
A Scheduler that lowers license and server costs for its users and improves the scalability of their IT infrastructure
An explainability module that uses a Large Language Model to explain the decisions made by the Scheduler
The project FW11020080 is co-financed with state support from the Technology Agency of the Czech Republic.
Highlights
- Scheduling of Robotic Process Automation tasks.
- Lower license and server costs of RPA infrastructure.
- Explainable scheduling decisions using Large Language Models.
Team
- Přemysl Šůcha – Principal investigator
- Vilém Heinz – Problem analysis, scheduling algorithms, LLM-based explanations